Characterizing the relationship between temperature and mortality in tropical and subtropical cities: a distributed lag non-linear model analysis in Hue, Viet Nam, 2009-2013

• Published in: Global health action Vol. 9; no. 1; p. 28738
• Main Authors: Dang, Tran Ngoc, Seposo, Xerxes T., Duc, Nguyen Huu Chau, Thang, Tran Binh, An, Do Dang, Hang, Lai Thi Minh, Long, Tran Thanh, Loan, Bui Thi Hong, Honda, Yasushi
• Format: Journal Article
• Language: English
• Published: United States Taylor & Francis 01.01.2016; Taylor & Francis Ltd; Co-Action Publishing; Taylor & Francis Group
• Subjects: Adolescent; Adult; Age Distribution; Aged; Cardiovascular diseases; Child; Child, Preschool; Cities; Climate change; cold effects; Female; Global Health; High temperature; high temperature effects; hot effects; Humans; Infant; Infant, Newborn; Linear analysis; low temperature effects; Male; Middle Aged; Mortality; Nonlinear Dynamics; Older people; Original; Preventive medicine; Public health; Respiratory diseases; Risk Factors; Sex Distribution; Studies; Temperature; Temperature effects; time-series regression; Vietnam - epidemiology; Weather; Young Adult; Vietnam; Hue Vietnam; time-series regression; hot effects; high temperature effects; low temperature effects; cold effects
• ISSN: 1654-9716; 1654-9880
• DOI: 10.3402/gha.v9.28738

The relationship between temperature and mortality has been found to be U-, V-, or J-shaped in developed temperate countries; however, in developing tropical/subtropical cities, it remains unclear. Our goal was to investigate the relationship between temperature and mortality in Hue, a subtropical city in Viet Nam. We collected daily mortality data from the Vietnamese A6 mortality reporting system for 6,214 deceased persons between 2009 and 2013. A distributed lag non-linear model was used to examine the temperature effects on all-cause and cause-specific mortality by assuming negative binomial distribution for count data. We developed an objective-oriented model selection with four steps following the Akaike information criterion (AIC) rule (i.e. a smaller AIC value indicates a better model). High temperature-related mortality was more strongly associated with short lags, whereas low temperature-related mortality was more strongly associated with long lags. The low temperatures increased risk in all-category mortality compared to high temperatures. We observed elevated temperature-mortality risk in vulnerable groups: elderly people (high temperature effect, relative risk [RR]=1.42, 95% confidence interval [CI]=1.11-1.83; low temperature effect, RR=2.0, 95% CI=1.13-3.52), females (low temperature effect, RR=2.19, 95% CI=1.14-4.21), people with respiratory disease (high temperature effect, RR=2.45, 95% CI=0.91-6.63), and those with cardiovascular disease (high temperature effect, RR=1.6, 95% CI=1.15-2.22; low temperature effect, RR=1.99, 95% CI=0.92-4.28). In Hue, the temperature significantly increased the risk of mortality, especially in vulnerable groups (i.e. elderly, female, people with respiratory and cardiovascular diseases). These findings may provide a foundation for developing adequate policies to address the effects of temperature on health in Hue City.